Research On Robust Predictive Control Method For Modular Multi-winding Pmsm System Considering Parameter Mismatch

Facing on the new round of global industry transformation and industrial revolution,the "Made in China 2025" plan clearly proposes to vigorously promote breakthrough development in key areas such as advanced rail transportation equipment,advanced numerical control machine,new energy vehicles and aerospace equipment.However,high performance permanent magnet motor system is the key and core of high-end technical equipment in the above key areas.The shutdown of permanent magnet motor system will bring fatal impact to industrial equipment,and it will lead to catastrophic accidents in serious cases.Therefore,it is of great theoretical and practical significance to carry out the research on the design and highperformance control method of new multi-phase permanent magnet motor in some special application fields with higher and stricter reliability requirements,so as to promote the highend technology and equipment in key fields of our country to be in line with the international advanced level.Aiming at the key common bottleneck problem of high performance control for a new modular multi-winding permanent magnet motor system,this dissertation explores the design and high performance control of multi-winding permanent magnet motor on the basis of major technical equipment development project.Based on the traditional design and control methods of permanent magnet motor,the modular multi-winding permanent magnet motor system design and high-performance control are explored,and a high-performance robust predictive control method suitable for the new modular multi-winding permanent magnet motor is developed in this dissertation.The research contents include: modular multi-winding permanent magnet motor ontology design,characteristic analysis,operation mechanism and model establishment;parameter mismatch sensitivity analysis of modular multi-winding permanent magnet motor control performance;and parameter mismatch robust predictive control method of modular multi-winding permanent magnet motor.A relatively perfect system design and control technology system of modular multi-winding permanent magnet motor is established,which provides a certain theoretical guidance and reference value for engineering practice.The innovation and key research contents of this dissertation include:(1)The finite element simulation model of the new modular multi-winding permanent magnet motor is established.The output characteristics and operation mechanism of the motor are analyzed by finite element calculation.On this basis,the accurate mathematical model of the modular multi-winding permanent magnet motor is obtained,which lays the theoretical foundation for the high performance control of the modular multi-winding permanent magnet motor system.According to the special winding structure and characteristics of the new modular multi-winding permanent magnet motor,the direct power predictive control method and deadbeat predictive control method are proposed in this dissertation.The proposed method eliminates the influence of irreversible one beat delay in the digital control system,effectively suppresses the oscillation problem caused by the high bandwidth predictive speed control,meets the high performance speed regulation requirements of the modular multi-winding permanent magnet motor system,realizes the low torque pulsation operation of the modular multi-winding permanent magnet motor,and expands the reliability of the permanent magnet motor in harsh environment.It is suitable for the application in the situation of high requirement of the security.(2)The inherent nature of parameter mismatch affecting the control performance of permanent magnet motor system is revealed,and the sensitivity of model predictive controller to parameter mismatch is analyzed in detail.The novel predictive stator flux control method is then proposed in this dissertation.A composite discrete sliding mode observer based on stator flux state is designed,which can estimate flux parameter perturbation,rotor position error and load torque disturbance simultaneously.The composite discrete sliding mode observer is used to compensate the predictive controller,which can eliminate the influence of flux parameter perturbation and rotor position error on the predictive controller.The control performance of traditional predictive control method and novel predictive stator flux control method is compared.The results show that the proposed novel predictive stator flux control method can achieve low stator current harmonic and low torque ripple operation under the condition of flux parameter perturbation and rotor position error,and improve the control performance of motor system under parameter mismatch and rotor position error.(3)The robust torque predictive control method of modular multi-winding permanent magnet motor is studied for high-power traction applications.A modular multi-winding permanent magnet motor system driven by six parallel inverters is designed in this dissertation.Based on the conventional predictive torque control method,the parameter mismatch sensitivity of resistance,inductance and flux linkage to the control performance of modular multi-winding permanent magnet motor system is analyzed.By introducing a proportional controller into the torque predictive controller,the control accuracy and robustness of the modular multi-winding permanent magnet motor system are effectively improved under parameter mismatch.The control performance of modular multi-winding permanent magnet motor system with conventional torque predictive control method is compared under parameter mismatch.The results show that the robust torque predictive control method can significantly reduce the stator flux and torque ripple under parameter mismatch.(4)The parameter immune predictive control method of modular multi-winding permanent magnet motor is proposed and studied in this dissertation.The robust stator flux predictive controller based on one beat delay compensation is designed by analyzing the robustness of stator flux predictive controller to parameter mismatch,which can effectively suppress the influence of parameter mismatch and one beat delay on the control performance of modular multi-winding permanent magnet motor system.At the same time,a robust torque predictive controller based on unknown torque disturbance observer is designed,which can improve the robustness to load torque disturbance and parameter mismatch.The transient performance and steady-state performance of conventional torque predictive control method and parameter immune predictive control method are compared under parameter mismatch.The results show that parameter immune predictive control method has obvious advantages in improving stator flux tracking accuracy,restraining torque/flux ripple and reducing stator current distortion.This dissertation focuses on the major common needs of China's manufacturing industry for the innovation and development of high-end technology and equipment.Taking the permanent magnet motor system widely used in modern manufacturing industry as the research object,this dissertation focuses on the robust predictive control method of the new modular multi-winding permanent magnet motor system.The common basic problem of high performance control of modular multi-winding permanent magnet motor system is solved under parameter mismatch.A relatively perfect design,optimization and control theory of modular multi-winding permanent magnet motor is formed in this dissertation.It expands the breadth and depth of permanent magnet motor system in different application fields of manufacturing industry,and provides theoretical guidance and reference for promoting the engineering practice of modular multi-winding permanent magnet motor system.